Abstract Deep brain stimulation (DBS) has a major role in the management of movement disorders, and is under investigation for the treatment of disorders of mood and memory. In Parkinson's disease (PD), DBS of basal ganglia nuclei can improve motor signs and reduce medication-induced motor fluctuations and dyskinesia, characterized by frequent transitions between a hypokinetic state (too little movement) and a hyperkinetic state (too much movement). However, since the introduction of DBS for PD 25 years ago, there have been no major improvements in this therapy. Existing DBS devices deliver ?open loop? stimulation, continuously stimulating their target structures regardless of changes in the brain circuits related to disease expression. Device programming is a labor-intensive process based on ?trial and error? requiring significant clinical expertise, which is a barrier to widespread application. In PD, continuous open-loop stimulation may result in suboptimal control of fluctuating motor signs, stimulation-induced adverse effects, and short battery life. DBS could be significantly improved by delivering ?closed-loop? stimulation, in which stimulation parameters are automatically adjusted based on brain signals that reflect the patient's clinical state. Using both intraoperative and chronic invasive recording techniques, we and others have identified abnormal patterns of oscillatory activity that may provide physiological signatures or ?biomarkers? of hypokinetic and hyperkinetic states in PD. Here, we plan to develop closed-loop DBS algorithms based on these brain signals, using an investigational neural interface (Medtronic Activa RC+S) that can sense and store brain activity as well as delivering DBS. We will determine which brain signals are the most appropriate to optimize DBS therapy and answer critical questions including the site of control signal detection (cortical versus subcortical) and the required complexity of control signals (single frequency power versus cross frequency interactions). Ten PD patients with motor fluctuations and dyskinesia will be implanted bilaterally with Activa RC+S attached to a subthalamic nucleus (STN) DBS lead and an electrocorticography (ECoG) lead placed over motor cortex. We will collect ECoG and subcortical local field potential (LFP) recordings to characterize ?personalized? physiological signatures for each subject and prototype stimulation paradigms by data streaming through an external computer in a clinical setting (Aims 1 and 2). We will then embed algorithms in the pulse generator to implement chronic and fully closed-loop DBS in a small double-blinded clinical trial (Aim 3). Motor function will be assessed by wearable automated detectors as well as rating scales from videotapes and self-report instruments. The study will define the technical characteristics required for the design of future DBS devices. ?Self programming? DBS devices offer the potential to simplify the therapy and allow many more patients to receive DBS. As electrophysiological signatures of abnormal circuit function in other disorders are identified, this new generation of closed-loop devices will facilitate the introduction of novel DBS therapies in other brain diseases. .